gprof with gnat

gprof with gnat

[Charles Fr. Rey]

Anyone knows how to use gprof with gnat (linux version) ? I was able to
use it on solaris, but not on my home computer ...

Thanks in advance !

Re: gprof with gnat

[Eric G. Miller]

In <3CCD2328.C5722FC3@epfl.ch>, Charles Fr. Rey wrote:

> Anyone knows how to use gprof with gnat (linux version) ? I was able to
> use it on solaris, but not on my home computer ...

AFAIK, it doesn't work on Ada (says C, Pascal, and Fortran77 in the manual).

Re: gprof with gnat

[Charles Fr. Rey]

Yes, but it works on solaris, adding a parameter to the linker (link
with gmon.o). So why not on Linux ? (I don't find any gmon.o that comes
with gnat package for linux) ...

"Eric G. Miller" wrote:
> 
> In <3CCD2328.C5722FC3@epfl.ch>, Charles Fr. Rey wrote:
> 
> > Anyone knows how to use gprof with gnat (linux version) ? I was able to
> > use it on solaris, but not on my home computer ...
> 
> AFAIK, it doesn't work on Ada (says C, Pascal, and Fortran77 in the manual).

Re: gprof with gnat

[Charles Fr. Rey]

Is Ada so performant that it doesn't need to be profiled ? ;-)

Re: gprof with gnat

[Ted Dennison]

"Charles Fr. Rey" <charlesfr.rey@epfl.ch> wrote in message news:<3CCE6DF1.46B5450E@epfl.ch>...
> Is Ada so performant that it doesn't need to be profiled ? ;-)

Not really.

I attempted to use it on Windows about a year ago, with rather rotten
luck. Most of the problems stem from the fact that Gnat doesn't use
the latest gcc, and thus you can't use the latest gprof. Hopefully
this problem will go away with gcc 3.1.


-- 
T.E.D. 
Home     -  mailto:dennison@telepath.com (Yahoo: Ted_Dennison)
Homepage -  http://www.telepath.com/dennison/Ted/TED.html

Re: gprof with gnat

[Gautier]

Charles Fr. Rey:

> > Is Ada so performant that it doesn't need to be profiled ? ;-)

T.E.D.:
 
> Not really.
> 
> I attempted to use it on Windows about a year ago, with rather rotten
> luck. Most of the problems stem from the fact that Gnat doesn't use
> the latest gcc, and thus you can't use the latest gprof. Hopefully
> this problem will go away with gcc 3.1.

What are the problems ?
At least my small experiments with gprof (provided with GNAT 3.13p for
Windows) do run well...
E.g. by profiling the Unzip-Ada tool (trying to analyse the slow
I/Os).
___________________________________________________________
Gautier  --  http://www.mysunrise.ch/users/gdm/unzipada.htm

NB: For a direct answer, address on the Web site!

Re: gprof with gnat

[Ted Dennison]

gautier_niouzes@hotmail.com (Gautier) wrote in message news:<17cd177c.0204301131.63213cc2@posting.google.com>...
> > I attempted to use it on Windows about a year ago, with rather rotten
> > luck. Most of the problems stem from the fact that Gnat doesn't use
> > the latest gcc, and thus you can't use the latest gprof. Hopefully
> > this problem will go away with gcc 3.1.
> 
> What are the problems ?
> At least my small experiments with gprof (provided with GNAT 3.13p for
> Windows) do run well...
> E.g. by profiling the Unzip-Ada tool (trying to analyse the slow
> I/Os).

It took quite a while to get hold of a properly linked gprof, as most
of the ones out there use the newer version of gcc's libraries, which
are incompatable with the current public versions of gnat (or at least
were at the time). It took quite a bit of searching to find the proper
gprof. That was the hard part.

I finally did get some useful results, but as I remember some of the
gprof options were not working properly (I'd get no data). Of course
profilers are notoriously tricky to use, so I could easly have been
*correctly* getting no data. But I was able to get data with the same
options on Solaris...


-- 
T.E.D. 
Home     -  mailto:dennison@telepath.com (Yahoo: Ted_Dennison)
Homepage -  http://www.telepath.com/dennison/Ted/TED.html

Re: gprof with gnat

[Robert Dewar]

dennison@telepath.com (Ted Dennison) wrote in message news:<4519e058.0204300558.49ff7b08@posting.google.com>...
> "Charles Fr. Rey" <charlesfr.rey@epfl.ch> wrote in message news:<3CCE6DF1.46B5450E@epfl.ch>...
> > Is Ada so performant that it doesn't need to be profiled ? ;-)
> 
> Not really.
> 
> I attempted to use it on Windows about a year ago, with rather rotten
> luck. Most of the problems stem from the fact that Gnat doesn't use
> the latest gcc, and thus you can't use the latest gprof. Hopefully
> this problem will go away with gcc 3.1.

Be careful in reading posts on this group. The fact that Ted cannot
manage on his own to get something working does not mean that it does
not work :-), and the analysis in the second sentence is bogus and
incorrect.

In fact gprof works with GNAT on Windows on GCC 2
and we have several of our customers successfully using it.
Preliminary tests on GCC 3 seem to show that the behavior
is the same, but since we do not have a GCC 3 based product
yet, we do not have customers using gprof on GCC 3.

Robert Dewar
Ada Core Technologies

Re: gprof with gnat

[Ted Dennison]

dewar@gnat.com (Robert Dewar) wrote in message news:<5ee5b646.0205010430.4899201e@posting.google.com>...
> dennison@telepath.com (Ted Dennison) wrote in message news:<4519e058.0204300558.49ff7b08@posting.google.com>...
> > "Charles Fr. Rey" <charlesfr.rey@epfl.ch> wrote in message news:<3CCE6DF1.46B5450E@epfl.ch>...
> > > Is Ada so performant that it doesn't need to be profiled ? ;-)
> > 
> > Not really.
> > 
> > I attempted to use it on Windows about a year ago, with rather rotten
> > luck. Most of the problems stem from the fact that Gnat doesn't use
> > the latest gcc, and thus you can't use the latest gprof. Hopefully
> > this problem will go away with gcc 3.1.
> 
> Be careful in reading posts on this group. The fact that Ted cannot
> manage on his own to get something working does not mean that it does
> not work :-), and the analysis in the second sentence is bogus and
> incorrect.

I agree with the first half of this entirely. There are certianly
loads of people out there who know way more than me about building and
using gcc tools. The second part was certianly not true at the time.
If something was done to gcc 3 to  make non-EGCS executables work with
EGCS gprof executables, then I guess its not an issue any more. But a
year ago the EGCS gprof would *not* work with my gnat-built exes on
Windows. I found (after much searching) a non-EGCS gprof, and that
worked much better (but still not perfectly).


-- 
T.E.D. 
Home     -  mailto:dennison@telepath.com (Yahoo: Ted_Dennison)
Homepage -  http://www.telepath.com/dennison/Ted/TED.html

Re: gprof with gnat

[Bobby D. Bryant]

On Mon, 29 Apr 2002 04:40:40 -0600, Charles Fr. Rey wrote:

> Anyone knows how to use gprof with gnat (linux version) ? I was able
> to use it on solaris, but not on my home computer ...

I'm not very expert with gprof, but I tried it last fall and didn't
seem to have any problems.  I use Linux and GNAT.

Here are some notes I made at the time:


 o Comments in the source file for a jacket program, profile.adb, that
   I wrote to drive a GA "solver" for the profiling -

-- This is useful for testing for memory leaks in the solver:
--    gnatmake profile
--    gnatmem 5 ./profile
--
-- Also useful for optimizing the solver:
--    make profile
--    ./profile
--    gcov profile.adb
--    emacs profile.adb.gcov
--    n.b. -- you can also run gcov on files that profile.adb depends on.


 o A snippet from my Makefile -

# For use w/ the gcov profiler:
profile: realclean
        gnatmake -g profile                   \
        -cargs -fprofile-arcs -ftest-coverage \
        -bargs -C -largs -C


Hopefully that will be enough to get you started.

Good luck,

Bobby Bryant
Austi, Texas

Re: gprof with gnat

[Charles Fr. Rey]

gnatmake -g profile etc... it seems that's the missing thing ..  I was
used to use it on solaris, and you MUST add -largs path/gprof.o, no more
...

Thanks, next time i'll read the whole documentation !



"Bobby D. Bryant" wrote:
> 
> On Mon, 29 Apr 2002 04:40:40 -0600, Charles Fr. Rey wrote:
> 
> > Anyone knows how to use gprof with gnat (linux version) ? I was able
> > to use it on solaris, but not on my home computer ...
> 
> I'm not very expert with gprof, but I tried it last fall and didn't
> seem to have any problems.  I use Linux and GNAT.
> 
> Here are some notes I made at the time:
> 
>  o Comments in the source file for a jacket program, profile.adb, that
>    I wrote to drive a GA "solver" for the profiling -
> 
> -- This is useful for testing for memory leaks in the solver:
> --    gnatmake profile
> --    gnatmem 5 ./profile
> --
> -- Also useful for optimizing the solver:
> --    make profile
> --    ./profile
> --    gcov profile.adb
> --    emacs profile.adb.gcov
> --    n.b. -- you can also run gcov on files that profile.adb depends on.
> 
>  o A snippet from my Makefile -
> 
> # For use w/ the gcov profiler:
> profile: realclean
>         gnatmake -g profile                   \
>         -cargs -fprofile-arcs -ftest-coverage \
>         -bargs -C -largs -C
> 
> Hopefully that will be enough to get you started.
> 
> Good luck,
> 
> Bobby Bryant
> Austi, Texas

Re: gprof with gnat

[Robert Dewar]

"Charles Fr. Rey" <charlesfr.rey@epfl.ch> wrote in message news:<3CCF8702.A7537851@epfl.ch>...
> gnatmake -g profile etc... it seems that's the missing thing ..  I was
> used to use it on solaris, and you MUST add -largs path/gprof.o, no more
> ...
> 
> Thanks, next time i'll read the whole documentation !

<<followed by dozens of lines quoting the entire original message>>

Please don't quote whole messages in email or whole articles on a newsgroup.
It is abusive and careless. If you are using tools like Microsoft outlook
that let you turn this on by default TURN IT OFF! And take the trouble to
properly edit quotes. 

The worst case is when TWO such careless people talk to one another and
you get a quadratic explosion of quotes.

This problem seems to have got MUCH worse recently. Has there been some
annoying change in microsoft software?

Re: Bad reply practice (OT)

[Chad R. Meiners]

I don't think that it has become any worst than before.  Perhaps since you
ignore some threads in CLA, you are experiencing a higher hit ratio.
Anyway, you still make a valid point about the practice of including the
only the text being replied to.

-CRM

"Robert Dewar" <dewar@gnat.com> wrote in message
> This problem seems to have got MUCH worse recently. Has there been some
> annoying change in microsoft software?

Re: Bad reply practice (OT)

[Robert Dewar]

"Chad R. Meiners" <crmeiners@hotmail.com> wrote in message news:<aapbv1$2lv9$1@msunews.cl.msu.edu>...

> I don't think that it has become any worst than before.  Perhaps since you
> ignore some threads in CLA, you are experiencing a higher hit ratio.
> Anyway, you still make a valid point about the practice of including the
> only the text being replied to.

Probably it is because more people are using tools that automatically quote
stuff. I know for example that many of the news readers that I used to use (I
now use Google) would monitor quotes and question or reject posts with an
excessive ratio of quotes to text.

For me, a reasonable guideline is that a post should never have more quoted
text than non-quoted text. Far more than fifty percent of messages on CLA
from a little statistical sample I did just now violate this rule.

We have threaded messages here. Furthermore, messages almost always have a
message identifier at the top (as this one does), so all you have to do in
the news group context is to quote enough to make it clear what you are 
responding to. If someone wants to see the entire message you are responding
to they can easily do so.

Re: gprof with gnat

[Charles Fr. Rey]

I use netscape/UNIX ... :-), so it's not a cro$oft problem ...
there's a "Automatically quote original message when replying" option,
but I guess that if I turn it off, the original message will still be
there but without quotes (and thus more unreadable) ... if you have a
better solution, tell me, please.

Re: gprof with gnat

[Robert Dewar]

"Charles Fr. Rey" <charlesfr.rey@epfl.ch> wrote in message news:<3CD0E169.88A260A0@epfl.ch>...
> I use netscape/UNIX ... :-), so it's not a cro$oft problem ...
> there's a "Automatically quote original message when replying" option,
> but I guess that if I turn it off, the original message will still be
> there but without quotes (and thus more unreadable) ... if you have a
> better solution, tell me, please.

You guess wrong, turn it off!

Re: [OT] ettiquite - was gprof with gnat

[Mark Johnson]

Robert Dewar wrote:
> 
> "Charles Fr. Rey" <charlesfr.rey@epfl.ch> wrote in message news:<3CD0E169.88A260A0@epfl.ch>...
> > I use netscape/UNIX ... :-), so it's not a cro$oft problem ...
> > there's a "Automatically quote original message when replying" option,
> > [snip]
> 
> You guess wrong, turn it off!

[flame on]
Please calm down - I think there's an intolerance virus going around
that causes people to get nasty about proper net etiquette. I see this
kind of response on several newsgroups and believe it is unnecessary. I
suggest a personal note to the OP instead of a public one if you feel a
rebuke is necessary. That I find is an UN common courtesy that should be
present. This rebuke is public to make the point to others (I'll never
change you Robert :-).
[flame off]

Back to the OP. I use Netscape for news as well, but you can see my
posts properly annotated. You can leave the quoting option enabled - to
make it more clear who said what, but take the time to cut the part you
don't need. Put something like [snip] in if you feel uncomfortable to
quote someone else's material in pieces. It also helps if your news
server requires more original text than quoted text - one server I use
regularly enforces this so I got into the habit of short quotes.

  --Mark

Re: [OT] ettiquite - was gprof with gnat

[Ted Dennison]

Mark Johnson <mark_h_johnson@raytheon.com> wrote in message news:<3CD14A98.6CF0B177@raytheon.com>...
> posts properly annotated. You can leave the quoting option enabled - to
> make it more clear who said what, but take the time to cut the part you
> don't need. Put something like [snip] in if you feel uncomfortable to

To me this is something you should *want* to do anyway. If there's
tons of unrelated cruft in your message, a lot of people will just
bail out without bothering to read what you are trying to say. Unless
you are just here for the typing practice, you should be interested in
maximizing the impact of your messages.

I know I always get upset with myself when I see extra quoted text I
forgot to take out.


-- 
T.E.D.
Home     -  mailto:dennison@telepath.com (Yahoo: Ted_Dennison)
Homepage -  http://www.telepath.com/dennison/Ted/TED.html

Re: [OT] ettiquite - was gprof with gnat

[Robert Dewar]

Mark Johnson <mark_h_johnson@raytheon.com> wrote in message news:<3CD14A98.6CF0B177@raytheon.com>...
> [flame on]
> Please calm down - I think there's an intolerance virus going around
> that causes people to get nasty about proper net etiquette. I see this
> kind of response on several newsgroups and believe it is unnecessary. I
> suggest a personal note to the OP instead of a public one if you feel a
> rebuke is necessary. That I find is an UN common courtesy that should be
> present. This rebuke is public to make the point to others (I'll never
> change you Robert :-).
> [flame off]

"please turn it off" hardly seems to be violent intolerance, and besides,
I note that the above note is a nice example of not following your own
advice :-) The reason that I think it is important to complain about excessive
quoting is that it is one of the things that greatly increases noise to
signal ratio in a newsgroup. Comp.lang.ada used to be far more useful than
it is, and it is a pity to see how many people have abandoned it. In this
case we had someone who claimed that if they turned off the quoting option,
they assumed that this meant that the message would be included without 
quotes. That's plain misinformation that is worth correcting, since otherwise
someone else might make the same mistake.

The reply to this message has it right. If you people to read your messages,
be more careful quoting junk. particularly annoying, and increasingly more
common is two people talking to one another quoting entire messages, and 
amazingly these programs have no guard against recursive quoting of messages
so you get quadratic accumulation of junk. I have seem messages quoted four
levels deep :-(

I find that I bother to look at fewer and fewer CLA threads these days, because
so many of them have excessive noise to signal ratio.

Re: [OT] ettiquite - was gprof with gnat

[Hyman Rosen]

Robert Dewar wrote:
> I find that I bother to look at fewer and fewer CLA threads these days,
 > because so many of them have excessive noise to signal ratio.

I think you have been saying this ever since I can remember :-)

Re: [OT] ettiquite - was gprof with gnat

[Robert Dewar]

Hyman Rosen <hyrosen@mail.com> wrote in message news:<3CD28EF8.40004@mail.com>...
> Robert Dewar wrote:
> > I find that I bother to look at fewer and fewer CLA threads these days,
>  > because so many of them have excessive noise to signal ratio.
> 
> I think you have been saying this ever since I can remember :-)


Actually I used to follow all threads until deja closed down. Then I
stayed off CLA completely for a while, and
realized that it was not a great loss :-)

Since google reinstatated things, I have been steadily
reading fewer and fewer threads, so your memory is not
that long but indeed I have said this more than once
(and it is true :-)

Re: [OT] ettiquite - was gprof with gnat

[Bobby D. Bryant]

On Fri, 03 May 2002 07:22:00 -0600, Hyman Rosen wrote:

> Robert Dewar wrote:
>> I find that I bother to look at fewer and fewer CLA threads these
>> days, because so many of them have excessive noise to signal ratio.
> 
> I think you have been saying this ever since I can remember :-)
 

I think there are monotonic increases in noise and decreases in signal
in _all_ newesgroups.  Presumably he can continue saying that for ever.

Bobby Bryant
Austin, Texas

p.s. - ...suggesting something analogous to the second law of
thermodynamics for Usenet...

Re: [OT] ettiquite - was gprof with gnat

[Robert Dewar]

"Bobby D. Bryant" <bdbryant@mail.utexas.edu> wrote in message news:<ab8qnt$2d1$1@geraldo.cc.utexas.edu>...
> p.s. - ...suggesting something analogous to the second 
> law of thermodynamics for Usenet...


chuckle chuckle :-)

There's fun to be found in CLA after all :-)

Robert

Re: gprof with gnat

[Charles Fr. Rey]

OK for the option .. I'll try it ..
BTW, I think it's sometime necessary to leave the original message
quoted, so if you save it or whatever else, you got the whole thing in
one message, and you don't need to read the 15 levels of previous
messages "manually". I also agree that unnecessary text in a reply
should be removed.
For example, this message is a perfect message (in the form): reply
message on top, only two levels of quotes, and a newly arrived ngreader
can understand what we're talking about. Don't you agree ?

Robert Dewar wrote:
> 
> "Charles Fr. Rey" <charlesfr.rey@epfl.ch> wrote in message news:<3CD0E169.88A260A0@epfl.ch>...
> > I use netscape/UNIX ... :-), so it's not a cro$oft problem ...
> > there's a "Automatically quote original message when replying" option,
> > but I guess that if I turn it off, the original message will still be
> > there but without quotes (and thus more unreadable) ... if you have a
> > better solution, tell me, please.
> 
> You guess wrong, turn it off!

Re: gprof with gnat

[Larry Kilgallen]

In article <3CD237A7.B91E6F32@epfl.ch>, "Charles Fr. Rey" <charlesfr.rey@epfl.ch> writes:

> BTW, I think it's sometime necessary to leave the original message
> quoted, so if you save it or whatever else, you got the whole thing in
> one message, and you don't need to read the 15 levels of previous
> messages "manually". I also agree that unnecessary text in a reply
> should be removed.

That seems to be at least two (contradictory) viewpoints.  It
is not possible to leave something intact and also remove parts
of it.

> For example, this message is a perfect message (in the form): reply
> message on top, only two levels of quotes, and a newly arrived ngreader
> can understand what we're talking about. Don't you agree ?

Certainly not (if you only wanted Robert's response, I presume you
would have used email).  That to which you are responding should
_precede_ your response, not follow it.  If that seems like people
have to wade through too much to get to your comment, then you
are overquoting.

Re: gprof with gnat

[Robert Dewar]

Kilgallen@SpamCop.net (Larry Kilgallen) wrote in message news:<OX8Js6FlvLiU@eisner.encompasserve.org>...
> Certainly not (if you only wanted Robert's response, I 
> presume you would have used email).  That to which you 
> are responding should _precede_ your response, not follow > it.  If that seems like people have to wade through too 
> much to get to your comment, then you are overquoting.

Thanks Larry, a very good rule indeed!

Re: gprof with gnat

[Jean-Marc Bourguet]

"Charles Fr. Rey" <charlesfr.rey@epfl.ch> writes:

> For example, this message is a perfect message (in the form): reply
> message on top

A perfect message with a reply on the top.  Didn't you forget the
smiley?

Yours,

-- 
Jean-Marc

Re: gprof with gnat

[cfrey]

yes and no ... :-)
think about this : when you receive a send a reply letter to someone,
and you must send him a copy of the letter he previously wrote, you
put it in annex, right ? so I think it's possible to do the same with
messages in ng ... isn't it ?
but, i agree it's maybe not the common way to write messages ...

Jean-Marc Bourguet <jm@bourguet.org> wrote in message news:<3cd2816f$1@news.cadence.com>...
> "Charles Fr. Rey" <charlesfr.rey@epfl.ch> writes:
> 
> > For example, this message is a perfect message (in the form): reply
> > message on top
> 
> A perfect message with a reply on the top.  Didn't you forget the
> smiley?
> 
> Yours,

Re: gprof with gnat

[Stephen Leake]

chfr_rey@bluemail.ch (cfrey) writes:

> yes and no ... :-)
> think about this : when you receive a send a reply letter to someone,
> and you must send him a copy of the letter he previously wrote, you
> put it in annex, right ? 

Only because of the limitations of the paper medium. Cutting and
pasting paper would make the reply easier to read, but it is messy and
cumbersome. 

Or, if it's in a legal or formal context, keeping the original message
intact may be important, to lessen the chances of taking things out of
context. Newsgroups are very informal, and you can always get the
original anyway.

> so I think it's possible to do the same with messages in ng ...
> isn't it ? 

It is certainly possible; Marin David Condic (to pick one poster) does
it all the time.

However, I prefer mixed replies, like I'm doing here. It makes it much
clearer what you are responding to. When I read one of Marin's posts,
I first scroll to the bottom to see what he's responding to, then I
scroll back to the top (often a long way :) to read Marin's
contribution. Much better to present a linear flow.

> but, i agree it's maybe not the common way to write messages ...

Based on my personal survey of the newsgroups I read, the interspersed
reply style (as I am using here) is much more common. I think most
people feel it is more readable.

-- 
-- Stephe

Re: gprof with gnat

[Marin David Condic]

"Stephen Leake" <stephen.a.leake.1@gsfc.nasa.gov> wrote in message
news:ur8kp8isu.fsf@gsfc.nasa.gov...
>
> It is certainly possible; Marin David Condic (to pick one poster) does
> it all the time.
>
Just for the record, I usually go just by "Marin". I typically include the
"David" in correspondence because "Marin" might be ambiguous as to gender,
given that it is a rather unusual name and I'm not there to correct the
impression. :-) (The junk mail I get has had the name mangled just about
every conceivable way, so I don't get touchy about it. But there are times
life would be simpler if my parents had named me "Bill" or "George" :-)


> However, I prefer mixed replies, like I'm doing here. It makes it much
> clearer what you are responding to. When I read one of Marin's posts,

I will mix replies when I want to address individual points in a post - as
I'm doing now.


> I first scroll to the bottom to see what he's responding to, then I
> scroll back to the top (often a long way :) to read Marin's
> contribution. Much better to present a linear flow.
>
Everybody's tastes vary. For me, I've usually seen the original post and
want to get immediately to the reply. Scanning down through nested layers of
earlier posts to get to an "I Agree!" response is not what I want to do.
Hence I post on top - and have had some people jump on my case about it as
if it were a mortal sin. :-) My attitude is that the moment someone
purchases the newsgroup and has full and clear title to it, they can post
whatever rules they like for people to participate in it. In the mean time
they can possibly take some "toleration lessons" and I'll be their homework.
:-)


> > but, i agree it's maybe not the common way to write messages ...
>
What surprises me is how many people seem to take it so seriously. Relax
about it! Its only zeros and ones and any time they become inconvenient,
there's a delete key! Post on top. Post on the bottom. Post interspersed.
Quote a single line. Quote the last 12 layers of replies*. Quote nothing. If
you have something of value to say, the message will get across as long as
you don't burry it too deeply. Its more important to me that responses be
polite, courteous, respectful of other's feelings & egos, patient,
understanding, helpful, etc. Those are the things that make social
interaction work best. The format is just window dressing.

*I've got 10 gig of disk space available on my local drive and I know how to
delete/purge my news posts. I doubt seriously that anyone can quote enough
plain text in a newsgroup to make me notice. :-)


MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com

Re: gprof with gnat

[Stephen Leake]

"Marin David Condic" <dont.bother.mcondic.auntie.spam@[acm.org> writes:

> Everybody's tastes vary. For me, I've usually seen the original post and
> want to get immediately to the reply. Scanning down through nested layers of
> earlier posts to get to an "I Agree!" response is not what I want to do.

Agreed. That's precisely when snipping is in order :).

> Hence I post on top - and have had some people jump on my case about
> it as if it were a mortal sin. :-)

Not quite "mortal", but up there :).

> My attitude is that the moment someone purchases the newsgroup and
> has full and clear title to it, they can post whatever rules they
> like for people to participate in it. In the mean time they can
> possibly take some "toleration lessons" and I'll be their homework.
> :-)

Good point.

-- 
-- Stephe

Re: gprof with gnat

[Wes Groleau]

> think about this : when you receive a send a reply letter to someone,
> and you must send him a copy of the letter he previously wrote, you
> put it in annex, right ? so I think it's possible to do the same with

Not necessarily.  You don't generally need to send him
the entire original.  And you don't have to put it at the end.

> messages in ng ... isn't it ?

And a newsgroup is not a private communication, anyway.

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: gprof with gnat

[Charles Fr. Rey]

Wes Groleau wrote:
> 
> > think about this : when you receive a send a reply letter to someone,
> > and you must send him a copy of the letter he previously wrote, you
> > put it in annex, right ? so I think it's possible to do the same with
> 
> Not necessarily.  You don't generally need to send him
> the entire original.  And you don't have to put it at the end.

I agree .. but it was just an example, a new idea inspired from
real-life ... I just don't understand why some thing should be a rule
for everyone and not an another one. Did you know that there are many
ways to do the same thing, and not necessarily a better one ?


> 
> > messages in ng ... isn't it ?
> 
> And a newsgroup is not a private communication, anyway.

No ? please, tell me what it is ... I consider it is a kind of private
communication, like emails, but with more chance to get an answer, as
there's potentially more than one reader of your message. Alright, it's
not really private, but in a sense, yes, because the question or problem
or request you ask has always a private importance for you, hasn't it ?
If not, then you're just loosing your time ...

> 
> --
> Wes Groleau
> http://freepages.rootsweb.com/~wgroleau

WAY OFF TOPIC

[Wes Groleau]

> > And a newsgroup is not a private communication, anyway.
> 
> No ? please, tell me what it is ... I consider it is a kind of private

With e-mail, the recipient know what he sent and doesn't
need it quoted back, except maybe sometimes as a memory jogger.
With Usenet, the reader may have never seen the previous post.


Absolutely.  I do it all the time.

Gabriel Cadwallider posted:
blah, blah, yadda, yadda, for thirty lines and then
somewhere it says Has anyone ever done CPR? and follows
that by twenty more quoted lines, not to mention the
signature blocks of six prior posters......

How long would it take to figure out do _what_ all the time?

What percentage of people would bother to try?

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: WAY OFF TOPIC

[Preben Randhol]

On Tue, 07 May 2002 12:12:31 -0500, Wes Groleau wrote:
> 
>> > And a newsgroup is not a private communication, anyway.

Who said this?

>> 
>> No ? please, tell me what it is ... I consider it is a kind of private

And who said this?

> With e-mail, the recipient know what he sent and doesn't
> need it quoted back, except maybe sometimes as a memory jogger.
> With Usenet, the reader may have never seen the previous post.

Exactly, so please don't remove the name of the person(s) you are
quoting. :-)

Preben

Re: WAY OFF TOPIC

[Charles Fr. Rey]

Preben Randhol wrote:
> 
> On Tue, 07 May 2002 12:12:31 -0500, Wes Groleau wrote:
> >
> >> > And a newsgroup is not a private communication, anyway.
> 
> Who said this?
> 
> >>
> >> No ? please, tell me what it is ... I consider it is a kind of private
> 
> And who said this?
> 
> > With e-mail, the recipient know what he sent and doesn't
> > need it quoted back, except maybe sometimes as a memory jogger.
> > With Usenet, the reader may have never seen the previous post.
> 
> Exactly, so please don't remove the name of the person(s) you are
> quoting. :-)

:-)

Re: gprof with gnat

[Pascal Obry]

"Bobby D. Bryant" <bdbryant@mail.utexas.edu> writes:

> On Mon, 29 Apr 2002 04:40:40 -0600, Charles Fr. Rey wrote:
> 
> > Anyone knows how to use gprof with gnat (linux version) ? I was able
> > to use it on solaris, but not on my home computer ...
> 
> I'm not very expert with gprof, but I tried it last fall and didn't
> seem to have any problems.  I use Linux and GNAT.
> 
> Here are some notes I made at the time:
> 
> 
>  o Comments in the source file for a jacket program, profile.adb, that
>    I wrote to drive a GA "solver" for the profiling -
> 
> -- This is useful for testing for memory leaks in the solver:
> --    gnatmake profile
> --    gnatmem 5 ./profile
> --
> -- Also useful for optimizing the solver:
> --    make profile
> --    ./profile
> --    gcov profile.adb
> --    emacs profile.adb.gcov
> --    n.b. -- you can also run gcov on files that profile.adb depends on.
> 
> 
>  o A snippet from my Makefile -
> 
> # For use w/ the gcov profiler:
> profile: realclean
>         gnatmake -g profile                   \
>         -cargs -fprofile-arcs -ftest-coverage \
>         -bargs -C -largs -C

This is not gprof (profiling tool) arguments but gcov ones (a test coverage
tool).

Pascal.

-- 

--|------------------------------------------------------
--| Pascal Obry                           Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--|------------------------------------------------------
--|         http://perso.wanadoo.fr/pascal.obry
--|
--| "The best way to travel is by means of imagination"